Unlike other dimensions, data in parent entities is stored in the cube. So, you cannot have the consolidated values in a parent entity vary based on the user running the report.
The simplest and most performance efficient solution if you can't just add an intermediate parent in the existing hierarchy, as already mentioned, would be to have a second Entity hierarchy in which all the existing base entities roll up to a new parent and use PercentConsolidation=0 to exclude entities, or only map the existing base entities to be consolidated to the new parent. This solution requires a second hierarchy to be consolidated, but the only extra data generated is O#Elimination, O#Share between new parent & existing base, and the new parent. With this method base entities are only Calculated and Translated once, so the second consolidation will run quicker.
As also pointed out, you could retain a single Entity hierarchy, but copy the base entity data to another Scenario, or even use a hybrid or dynamic Scenario in which you exclude copying or set PercentConsolidation=0 for the entities to exclude. However, these options require all the base entity data to be duplicated one way or another, and a second complete consolidation. This is much more costly, as you are doubling the amount of data and consolidation will run longer since all the base entities need consolidating a second time.